I have uploaded bear v2.7.8 to SF
. Please note that, if you like speed and menu-driven mode as bear v2.7.7, you do not need to upgrade to v2.7.8. The new release has a graphic user interface (GUI) for setup and data manipulation. So it is slower than the menu-driven mode under Windows (not too bad if running under Ubuntu/Linux and iMac OS X.) If you like to watch its demo video, you can find it on Youtube
Major update logs
- Java installation: this step needs to be done first before all steps, if you have not installed any Java before. It is because we will install the package
rChoiceDialogs and this package requires
- Windows: go to Oracle to download JAVA, and select
Java for consumers to continue and follow the steps.
- iMac OS X: I really got confused when I worked with iMac OS X Sierra. What I did was to install Java obtained from Oracle first, this enabled the package
rChoiceDialogs be installed successfully. Then I run bear and the terminal showed that I need to install Java and brought me to a web page to download it. After installing the second Java, it run OK. Unfortunately, I don't remember that web page. It should be no problem for iMac users to figure out.
- Ubuntu/Linux: OpenJDK works well. Please go to browse this website for more details. Since I have Java installed on my system, I just do
sudo R CMD javareconf and it works fine.
- GTK drivers (and related files for GUI toolkit): if you are the first time to install bear, you need to do this. Please go and browse my previous post.
- installations of required packages: only do this step after the above two steps have finished successfully. The fastest way is to download
preinst.r from SF and place it in your R working directory. Finally, run it from R console/terminal by typing
source("preinst.r") and select
bear from the menu and type
0 to exist.
- Finally, you can install bear v2.7.8 now from a local zip file (Windows).
fread() from package
data.table to load all csv data files now;
fread() can recognize the file format of *.csv automatically.
- automatically saved manual selection of data points; no more ask users to save or not to save.
- fixed the errors of NCAplot(); thanks to Dr. Mahmoud Hassan Teaima (2016-07-28).
- added data check after loading from a csv file when doing NCA, like this, and probably this too.
- displayed plots on screen (device) again for Rgui/X11 or running with RStudio.
- added nonparametric statistics for Tmax for 2x2x2 crossover (please see this thread) as an optional selection. The output is like
--- statistical analysis of Tmax ---
Z-value* P-value* PE (T-R) Lower Upper
0.847 0.415 0.25 -0.250 0.750
*: Exact Wilcoxon-Mann-Whitney Test
alternative hypothesis: true mu is not equal to 0
-> Tmax has no statistical difference.
That's all. Sorry for this lengthy post and thanks for reading.